November 29, 2024
JJ says that Bittensor is "undervalued by 20 - 50X"
GM!
Permissionless III has been buzzing in Salt Lake City, with the crypto crowd hopping between panels, parties, and pickleball games.
But while everyone’s having fun, the crypto markets are on shaky ground. So, I have to ask:
Can we get back to building great products people want to use, please?
In this edition of our weekly AI Edge, we cover:
The Crypto AI market sits at $26.6B, marking a slight 0.8% dip from last week. Most crypto sectors stayed quiet, with no big moves to shake things up.
TAO is holding strong, now at $562.
Bittensor is making waves beyond the charts. Mogmachine, the founder of Taostats and Corcel, gave a talk at Google Toronto, which even caught the attention of YouTube co-founder Steve Chen, who reposted it! With mindshare on the rise, the market is taking note.
NMT and QUBIC both took hits of -18% and -10% respectively this week, likely feeling the hangover from their late-September rallies (~30%) as speculators dial back the risk.
A quick recap of the past 30-Days:
TAO, VIRTUAL, COMAI, SNS, ENQAI were the top 30D performers in our Crypto AI token watchlist
On the other hand, NOA, AR, AGIX, OCEAN and AKT were the worst 30D performers. Still, it’s not a bad showing since AKT is only down 2%.
The chart above tracks miner scores from Graphite AI’s Subnet 43.
And it’s clear: the miners are leaving commercial models in the dust for graph optimization. The gap is growing wider, with the green line at the top representing the miner models, while the blue line below is the commercial benchmark struggling to keep up.
Subnet 43 was built to tackle graphical problems with finesse. Think of it as the ultimate solver for the Traveling Salesman Problem (TSP)—a classic challenge of finding the shortest possible route through a set of cities and back again.
Potential use cases include planning travel itineraries or mapping out all the must-visit cafes in your area.
Source: GraphiteAI
The chart below shows miner models’ upward trajectory over time. In September, they outperformed commercial models by ~1.8%. By October, that lead stretched to ~2.5%.
Considering Graphite AI’s Subnet 43 has only been around for two months, this progress is impressive. They’ve scaled from tackling graph problems with just 250 nodes to handling challenges with 2,500–3,000 nodes. It’s a testament to how TAO incentives can motivate participants to reach new heights.
HISTORIC day for Ai Layer Labs and @Wire_Blockchain.
We announced our merger along with many other updates. 🤝
Check out @KenDicross “Freedom to Transact” speech below for updates 👇
— 6079.ai (@6079ai)
4:18 PM • Oct 10, 2024
There’s a buzz building around AI agents taking over the bulk of on-chain transactions. Kalos (Parallel, Wayfinder) is making a bold claim: within a year, 90% of all transactions could be agent-driven.
6079.ai or AI Layer Labs and Wire Network just announced their merger to make this future a reality.
6079.ai’s Proof of Inference Protocol (PoIP) aims to deliver robust verification and security for decentralized AI inference. Meanwhile, Wire Network is building a gasless, interoperable blockchain network tailored specifically for AI agents.
Both believe that AI agents will soon dominate on-chain activity and are laying the infrastructure to make it happen.
Their first big move? A universal DEX that lets AI agents trade assets across chains without gas fees. But how can a network run without gas fees, you ask?
Ken’s got the answer: it all comes down to network ownership. Unlike the usual setup where users buy gas tokens (think SOL or ETH), Wire Network is flipping the script. Instead, users can purchase nodes and set the gas prices for transactions flowing through them.
If a node owner whitelists a specific smart contract, they’ll cover the gas fees for users interacting with that contract.
Will we see more M&A activity in the Crypto AI space? Time will tell.
Fundraising
Bittensor Spotlight
ALPHA ALERT: if you want to stay up-to-date on the Bittensor ecosystem, TAO Times is our favourite sibling newsletter to follow!
Project Updates
Incentive / Rewards Programs
Arweave
It’s a big week for AI in the Nobel world.
The 2024 Nobel Prize in Physics went to John J. Hopfield and Geoffrey E. Hinton, recognizing their groundbreaking use of physics tools to lay the foundations for deep neural networks and modern day machine learning. Congrats to both of them!
But not everyone’s thrilled. There’s been some online chatter about whether they deserved the physics prize, given that their work leans more toward AI than traditional physics—leaving some other heavyweight contenders out in the cold.
Meanwhile, the Nobel Prize in Chemistry went to David Baker (for AI in computational protein design), Demis Hassabis (DeepMind co-founder) and John M. Jumper (AlphaFold creator)
Looks like even the Nobel committees are hopping on the AI hype train.
Fraction AI announces breakthrough to improve model performance
Today, I'm excited to share a breakthrough from our team at @FractionAI_xyz that could redefine AI development.
We've created Gamified Adversarial Prompting (GAP), a decentralized platform that crowdsources labeled data, specifically targeting model weaknesses. The key? We've… x.com/i/web/status/1…
— Shashank | Fraction AI (@0xshai)
5:53 AM • Oct 10, 2024
JJ lays out the bull case for Bittensor subnets
It’s a good question. Community size is exploding. Quality of developers is also increasing very rapidly (I spend a lot of my time on this). The incentive is huge and I think $TAO is currently undervalued by somewhere between 20-50X, with 1-5000X potential and I’m trying to be… x.com/i/web/status/1…
— JJ (@JosephJacks_)
7:01 PM • Oct 6, 2024
a16z: Building the AI brain
🚨 New @a16z thesis: building the "AI brain"
We all exist in our own context.
Is it possible to take your jumble of thoughts, history, and memories and distill it into something tangible?
How I exported my brain to AI + what I'm seeing from startups 👇
— Justine Moore (@venturetwins)
4:02 PM • Oct 10, 2024
A 250-page report on Edge AI
With Apple Intelligence's release just around the corner, Edge AI will be the biggest narrative heading into 2025.
We (@peri_labs) have written a 250-page report to showcase:
• The Need for Edge AI
• Big Ideas
• Core Frameworks
Key ideas are in this thread ↓
(1/9)
— advait jayant (@advaitjayant)
4:52 PM • Oct 10, 2024
That’s it for this week! If you have specific feedback or anything interesting you’d like to share, please just reply to this email. We read everything.
Cheers,
Teng Yan & Joshua
This newsletter is intended solely for educational purposes and does not constitute financial advice. It is not an endorsement to buy or sell assets or make financial decisions. Always conduct your own research and exercise caution when making investment choices
